Back to the changelog

• Feature

  Field registry and Data Schema tab

  Every flow now has a registry of its payload fields, with two souls. **Declared fields** are the contract: type, label and "Sensitive" flag (feeding the Data Shield), directly editable. **Observed fields** are those that actually appeared in payloads — written by forms, tasks, AI outcomes, n8n connectors or APIs — automatically recorded with detected type and provenance (channel and birth phase). One click on "Promote" turns an observed field into a declared one: the contract grows from real use. Existing flows populate themselves the first time you open the tab.

• Improvement

  Visible type conflicts, optional strict mode

  If someone writes a value whose type differs from the declared one, the value is accepted but the field shows a conflict badge with a count: no lost writes, no silent surprises. For flows that want rigour there is strict mode: non-conforming writes are rejected on every channel. The setting also travels with the flow's export.

• Design

  Flow settings reorganized

  The General tab is down to the essentials (name, description, code, folder, default operator, expected path); operational tuning — calendar rescheduling, default task policy, WIP limit, alert policy, forced status change — lives in the new **Advanced** tab. The "API Trigger" tab is gone: endpoint, preconfigured n8n node copy and API key management sit together in the API Trigger section of Activation, with no tabs appearing and disappearing.

• Improvement

  Form AI starts from your data structure

  The sample payload the AI uses to generate the start form now starts from the flow's declared fields instead of a fixed example. If the saved sample drifts from the schema, "Generate from Data Schema" realigns it in one click.